Prediction of hypertensive disorders in pregnancy based on placental growth factor

1.Introduction

Hypertensive disorders in pregnancy (HDP) is a group of diseases that coexist with pregnancy and hypertension [1, 2, 3]. It is one of the most common obstetric complications [4, 5]. In order to reduce the occurrence of HDP and adverse maternal and infant outcomes, the disease needs to be effectively predicted as early as possible. The background and factors of HDP are complicated. Many researchers mainly made predictions based on the characteristics of the mother, some biomarkers (sFlt-1 (Soluble fms-Like Tyrosine Kinase 1), PlGF (Placental Growth Factor), VEGF (Vascular Endothelial Growth Factor), et al.) and other aspects [6, 7, 8]. With the development of various aspects of technology and engineering, in addition to statistical prediction methods, there are also artificial intelligence algorithms such as machine learning. Sufriyana et al. used maternal characteristics, uterine artery (UtA) doppler measurement, sFlt-1 and PlGF in the second and third trimester of pregnancy to study the machine learning related model for predicting preeclampsia (PE) and intrauterine growth restriction [9]. Vascular biomarkers have been conducted many different types of comparative analysis [10]. In addition to simply using the value of biomarkers, there were cases where the ratio of sFlt-1/PlGF ratio was used for prediction. In 2016, the relevant prospective study was conducted in 14 European countries [11]. In 2019, the multi-center clinical study was jointly carried out in 25 Asian research centers to supplement Asian research data [12]. These studies were to verify the effectiveness of the sFlt-1/PlGF ratio in predicting the risk of PE in different populations. However, there are still controversies in some aspects about the research on biomarkers to predict HDP. Suresh et al. had doubts about the admission treatment of suspected pregnant women with PE by assessing angiogenic factors, and the clinical utility of these factors was unclear [13]. The predictive value of the sFlt-1/PlGF ratio for short-term twin pregnancy without PE and mother-infant or neonatal complications remained to be studied and verified [14]. Pluddemann and Annette found that the effect of PlGF blood examination for predicting PE still needed to continue to be evaluated for better and wider promotion [15].

There have been many studies on biomarkers, especially PlGF. But the domestic research is relatively less compared with foreign countries, and the clinical application is not explicit. PlGF serum is an invasive blood test. Compared with noninvasive parameters, obstetricians will use PlGF more cautiously. Biomarkers such as PlGF lack relevant research on the clinical outcome of mothers and infants, and are not widely used in domestic clinical examinations. Biomarkers and other maternal parameters need to be jointly applied to predict HDP. It is also necessary to combine clinical research evidence from more Chinese populations to formulate a plan that meets China’s national conditions.

2.Materials and methods

3.Results

Incorporating the above-mentioned factors into the logistic regression equation to obtain two results with and without PlGF were as follows:

Through the logistic regression equation, the incidence probability P could be calculated, which could intuitively reflect the risk of HDP in pregnant women. ROC curve analysis was performed according to the incidence probability P. The comprehensive results of the two models are shown in Table 4.

Comprehensive indicators such as AUC and AC are more suitable for comprehensive analysis of model quality. According to Table 4, the AUC and AC of the model with PlGF were higher than those of the model without PlGF. In particular, the AC of the model with PlGF was above 90%. Therefore, adding the PlGF factor to the model for predicting HDP could make the model better.

4.Discussion

Lacerda et al. used VEGF, PlGF and sFlt-1 serum levels to distinguish active lupus nephritis and preeclampsia during pregnancy [16]. Black et al. used a multivariate screening algorithm for joint screening in the second trimester [17]. Maternal factors, mean arterial pressure (MAP), mean uterine artery pulsatility index (UtAPI), serum PlGF level in multiples of the median (MoM), and sFlt-1 level in MoM were used to predict premature delivery preeclampsia Combining information such as biomarkers (PlGF), Stepan et al. could improve the first trimester prediction and preeclampsia diagnosis [18].

In the aspect of data processing standardization, this research conducted qualitative and unified analysis. It was more in line with the results of the binary logistic regression model. Maternal basic factors combined with the biomarker (PlGF) for HDP comprehensive prediction was helpful to improve the accuracy and quality of the model. The discriminant effect was better than the prediction of a single class without PlGF.

5.Conclusion

According to the current domestic clinical needs and data, studying the role of PlGF as a biomarker. The actual role of PlGF was mainly researched in the overall prediction model. Simultaneously, combined with clinical needs and other basic maternal parameters for prediction. Based on the overall model, the comparison was made with or without PlGF. Adding the PlGF factor to the model for predicting HDP could improve the accuracy and effectiveness of the model. This study confirmed the predictive value and performance of PlGF. The amount of data and types of factors need to be increased in the future. And the data of people with HDP with complications should continue to be studied.